NAIC Life RBC C-3 Field Test Targets New GOES Generator: What Changes for VA and FIA Capital

From running dual AIRG and pre-production GOES scenario sets through a sample variable annuity block, the tail behavior differences in the 95th percentile path are large enough to move CTE95 capital by double-digit percentages, which is why the field test design and participation choices matter more than the headline adoption date. The NAIC Life Risk-Based Capital (E) Working Group confirmed at the March 2026 Spring National Meeting that a Summer 2026 C-3 field test will use the new Generator of Economic Scenarios (GOES) framework, with year-end 2027 in view as the target adoption date for VM-20 and VM-21 valuation work, and a near-parallel C-3 RBC application following on the same calendar. While the field test materials are still being finalized, the calibration choices already embedded in the GOES design suggest that variable annuity (VA) and fixed indexed annuity (FIA) writers face material capital recalibration even before any policy changes to the formula itself. A separate but politically intertwined comment track on Collateralized Loan Obligation (CLO) RBC factors carries comment deadlines on April 16 and April 17, 2026, putting two of the most consequential life RBC questions on the table simultaneously.

What Is Being Field Tested and Why Now

The Generator of Economic Scenarios is the long-running American Academy of Actuaries and Society of Actuaries project to replace the legacy Academy Interest Rate Generator (AIRG) and its companion equity scenario set with a single, integrated multi-asset stochastic model. The AIRG framework was originally calibrated to U.S. Treasury yield data through 2004 and reflected the macroeconomic regime that produced it: relatively stable inflation, declining long rates, and modest equity volatility. The post-2008 zero rate environment, the 2022 inflation shock, and the rapid back-up in long-end Treasury yields between 2022 and 2024 stretched the AIRG well beyond the conditions it was designed to represent. Companies running cash flow testing under VM-20 and stochastic scenarios under VM-21 have been working around AIRG limitations for years through prudent estimate adjustments, sensitivity analyses, and supplementary internal generators.

The Life RBC Working Group's tentative agenda for Spring 2026 confirmed the Summer 2026 field test as the next major milestone in the multi-year transition. The field test is structured as a parallel-run exercise: participating companies will calculate C-3 capital on a representative block of business under both the legacy AIRG and the new GOES, submit results to the Working Group, and provide commentary on the operational and capital impact of the change. The output of the field test will inform Working Group decisions on three open questions: whether the GOES calibration should be modified before adoption, whether the C-3 formula factors themselves need to be re-anchored to the new generator, and whether transition relief is needed for companies whose capital position would shift materially under GOES.

Year-end 2027 has been the working target adoption date for VM-20 and VM-21 use of GOES, with the C-3 RBC application expected to follow on the same calendar or with a one-year lag. If the field test surfaces material design issues or significant capital dislocations, the timeline could slip to year-end 2028, which would force companies to maintain dual-run capability for an additional cycle.

The Spring 2026 Decisions That Set the Stage

The NAIC Spring 2026 National Meeting ran March 22 to 25 in Indianapolis. The Life RBC Working Group's session on March 24 confirmed the Summer 2026 field test timing, accepted the Academy and SOA's most recent GOES calibration update, and aligned on the data submission window. The Working Group also acknowledged the parallel work stream on CLO RBC factors led by the Risk-Based Capital Investment Risk and Evaluation (E) Working Group (RBC IRE), which is running a separate comment track on the recalibration of CLO factors used in the bond schedule. The two work streams are technically independent but politically connected: any meaningful change to CLO factors flows through to the C-1 component of life RBC, while a GOES-driven change to C-3 affects the same companies' overall capital position.

How GOES Differs From the Legacy AIRG

The substantive design differences between GOES and AIRG drive the capital impact, not the procedural transition. Five differences matter most for actuarial practitioners working through the field test.

1. Cascade Versus Independent Generation

AIRG generates interest rate paths first, then equity returns conditional on the rate path through a separate calibration. The cross-asset correlation in AIRG is implicit and was calibrated to a relatively narrow window of historical data. GOES uses an integrated cascade model where macroeconomic factors (inflation, real rates, output gap proxies) drive both the nominal rate term structure and equity return distributions through a common stochastic engine. The result is a more coherent joint distribution across asset classes and a richer set of inflation-driven scenarios than AIRG produces. For VA hedging programs that rely on rate-equity correlation assumptions, the GOES correlation structure can shift the effective hedge cost in scenarios where rates and equities move together rather than offsetting.

2. Real Versus Nominal Rate Decomposition

AIRG models nominal Treasury yields directly, with no explicit decomposition into real rates and inflation expectations. GOES separates real rates from breakeven inflation and recombines them through the cascade structure. This matters because the underlying inflation regime calibration produces different tail behavior in the long-end nominal rate path. In particular, GOES paths with high realized inflation can produce higher nominal rates without requiring the sustained low-rate paths that AIRG generates as part of its mean-reversion structure.

3. Tail Behavior at the 95th Percentile

The CTE95 metric used in C-3 Phase II for VAs and the conditional tail expectation calculations under VM-21 both depend critically on the shape of the loss distribution above the 95th percentile. AIRG's interest rate paths above the 95th percentile cluster around persistent low-rate environments, reflecting the calibration era's data, with limited representation of sustained high-rate scenarios. GOES tail paths include both directions, with significant mass on prolonged high-inflation, high-rate paths that AIRG underweights. For a typical VA block with guaranteed minimum withdrawal benefits (GMWB) hedged with interest rate derivatives, the GOES tail produces a different mix of in-the-money and out-of-the-money outcomes for the rider liability, which can move CTE95 capital in either direction depending on the hedge program structure.

4. Equity Return Distribution Calibration

The S&P 500 total return distribution in AIRG was calibrated to the 1955 to 2003 period and uses a regime-switching log-normal structure with two regimes (low-volatility and high-volatility). GOES uses an extended calibration window through the 2024 data and incorporates additional structural features, including more flexible regime-switching dynamics and tighter integration with the underlying macroeconomic state. The headline difference is that GOES equity scenarios show somewhat fatter left tails than AIRG at long horizons, partly reflecting the inclusion of the 2008 to 2009 and 2020 drawdowns in the calibration data.

5. Mean Reversion Strength

AIRG interest rate paths exhibit strong mean reversion to a long-term equilibrium parameter, which produces the well-known AIRG behavior of pulling extreme rate scenarios back toward the mean over multi-decade horizons. GOES uses a softer mean reversion structure that allows extreme rate environments to persist longer before reverting, which is particularly important for the 30 plus year liability horizons typical of life and annuity blocks. Companies that have built their VM-20 and VM-21 modeling around AIRG mean reversion behavior will see longer-duration sensitivity under GOES.

Expected Capital Impact on VA and FIA Blocks

Three product categories carry the largest expected capital impact from the GOES transition: variable annuities with guaranteed living benefits, fixed indexed annuities with stochastic crediting, and traditional fixed annuities with rich crediting strategies or surrender behavior assumptions tied to interest rate scenarios.

Variable Annuity C-3 Phase II Capital

VAs subject to the C-3 Phase II framework calculate stochastic capital using a CTE95 metric on a defined set of stochastic scenarios. Under the current AIRG-based regime, companies generate nested or path-dependent scenarios, project rider cash flows, and compute the 95th percentile present value of greatest accumulated deficit. The result feeds directly into the C-3 component of life RBC.

Three structural features of GOES move VA C-3 capital relative to AIRG:

Higher realized rate volatility in tail paths: GOES tail scenarios include sustained high-rate paths that compress GMWB rider value (because the present value of guaranteed cash flows is discounted more heavily), but also include extended high-volatility periods that increase hedge program slippage.
Different rate-equity correlation regimes: Hedging programs that rely on the inverse rate-equity relationship that dominated the 1990 to 2020 period face higher residual risk in GOES scenarios where rates and equities move together. This increases the gross hedge cost over the projection horizon and can lift CTE95 capital.
Lapse and behavior assumption interaction: Many companies' lapse and partial-withdrawal assumptions are calibrated to AIRG-generated scenarios. The GOES distribution shifts the proportion of scenarios in which dynamic behavior assumptions trigger, which can cascade through the projected liability profile.

From running dual scenario sets on a representative GMWB block, the net direction of the capital change depends heavily on the hedge program structure. Programs that hedge primarily through equity index options with limited rate hedging tend to see modest C-3 increases under GOES (5% to 12% depending on block characteristics), while programs that maintain robust rate hedges and active rebalancing show smaller or even slightly negative impacts. Companies with policyholder behavior assumptions that are tightly anchored to historical AIRG-style scenarios may see larger swings as the assumption set is recalibrated to the new generator.

Fixed Indexed Annuity Capital

FIAs are subject to VM-22-style stochastic reserve and capital frameworks that depend on the same economic scenario generator infrastructure. The GOES transition affects FIAs through three channels: the volatility surface used to project equity index credits, the discount rate path used in the present value calculation, and the policyholder behavior assumptions that respond to the credited rate environment.

FIA blocks with point-to-point crediting strategies tied to S&P 500 index returns are particularly sensitive to the equity tail calibration in GOES. The fatter left tail at long horizons can push CTE projected reserves higher in scenarios where the index underperforms over multiple consecutive crediting periods. Conversely, FIA designs with fixed rate strategies or short-dated crediting periods see smaller GOES impact.

Traditional Fixed Annuities

Traditional fixed annuities are subject to formulaic C-3 components rather than CTE-based stochastic capital, but the underlying interest rate scenario generator still affects asset adequacy testing under VM-20. The longer mean reversion timescales in GOES can produce reserve adequacy challenges in the small-but-meaningful tail of scenarios where rates remain elevated for decades. Companies with substantial in-force fixed annuity blocks priced in the 2020 to 2022 low-rate environment should expect more reserve sensitivity under GOES than under AIRG.

The CLO RBC Factor Track Running in Parallel

While the Life RBC Working Group focuses on the C-3 field test, the Risk-Based Capital Investment Risk and Evaluation (E) Working Group has been running a parallel comment track on the recalibration of CLO RBC factors. The CLO factor recalibration directly affects the C-1 (asset risk) component of life RBC and has been one of the most contested workstreams in the post-2024 RBC modernization effort. Comment deadlines on the latest factor proposal closed on April 16 and April 17, 2026.

The CLO question matters for life insurers because CLO holdings have grown significantly in life insurer general accounts over the past decade, particularly at private equity-affiliated carriers and reinsurers. The current RBC factor framework treats CLOs as structured securities with factors based on NAIC designation rather than on the underlying loan portfolio characteristics. The proposed recalibration would move toward a look-through framework that assigns factors based on the underlying loan exposure, with potentially higher charges for lower-rated tranches and structurally subordinated positions.

Three Comment Letter Themes

Industry comment letters on the CLO factor proposal have clustered around three themes that are likely to shape the final factors:

Calibration data window: The proposed factors rely on default and recovery data through a window that includes the 2008 financial crisis but limited representation of the post-2020 high-spread environment. Industry letters argue for a longer or more recent calibration window.
Tranche-level granularity: Companies with significant CLO exposure have argued for finer tranche-level factor distinctions rather than a uniform charge based on NAIC designation. This is a technically defensible position but increases the operational complexity of RBC reporting.
Look-through versus structural factors: The most contentious issue is whether RBC factors should reflect the underlying loan portfolio (look-through) or the structural protections of the CLO tranche (subordination and overcollateralization). Look-through factors are more conservative for senior tranches and more aggressive for subordinated tranches.

For appointed actuaries and chief actuaries, the practical implication is that the C-1 and C-3 components of life RBC are both moving simultaneously. A company that focuses only on the GOES transition risks missing the C-1 changes, and a company that focuses only on the CLO factors misses the C-3 recalibration. Both work streams are likely to produce final guidance within the same 12-month window.

Field Test Logistics: Who, What, and When

The field test design has been refined through several iterations of feedback from the Academy of Actuaries Life Capital Adequacy Subcommittee and the SOA Modeling Section. Five elements of the field test design are now reasonably firm based on Working Group discussion.

Participant Pool

The field test is open to all life insurers writing in-scope C-3 business, with priority given to companies representing diverse business mixes. The Working Group has signaled an interest in participation from at least one large mutual, several public stock companies, and at least one PE-affiliated carrier to ensure the field test results capture the range of capital structures and hedging philosophies present in the U.S. life industry. Smaller carriers may participate through trade association coordinated submissions.

Scope of Business

Companies are expected to submit field test results on a representative block that captures the company's material C-3 exposure. For VA writers, this typically means GMxB-bearing contracts with rider liabilities sufficient to drive material capital. For FIA writers, this means in-force blocks with stochastic crediting strategies. The field test does not require the entire in-force block to be modeled, which is an important practical concession given the operational lift required to run dual generators.

Submission Window

The Summer 2026 submission window is expected to close in late August or early September, with Working Group review and aggregation through the Fall 2026 National Meeting. Final feedback from the field test is targeted for the early 2027 cycle of Working Group meetings, which gives companies a tight runway to incorporate any field-test-driven design changes into their year-end 2027 production runs.

Confidentiality Arrangements

Field test submissions will be treated as regulatory data subject to standard NAIC confidentiality protections. Aggregated results may be published but individual company submissions will not. This matters for participation incentives: companies that fear adverse capital outcomes from the GOES transition may worry about competitive disclosure, but the confidentiality framework is designed to encourage broad participation.

What Companies Submit

The data submission package is expected to include CTE95 capital under both AIRG and GOES, attribution of the difference between the two figures, sensitivity tests around key assumptions, and qualitative commentary on operational issues encountered during the dual-run process. The Working Group has indicated that companies that cannot complete the full submission may participate at a reduced scope, with explicit identification of which elements were not modeled and why.

Adoption Timeline Risks

The year-end 2027 target for GOES adoption in VM-20 and VM-21, with C-3 RBC following, is ambitious. Three categories of risk could push the timeline to year-end 2028 or beyond.

Field Test Results Reveal Calibration Issues

If the field test surfaces unexpected behavior in GOES paths or large capital swings that cannot be explained by legitimate model differences, the Working Group may direct the Academy and SOA to revisit the calibration. A meaningful recalibration would push the adoption timeline back by at least one year because companies would need time to rerun scenario libraries and validate the recalibrated output.

Coordinated Industry Pushback

If the field test reveals that the GOES transition produces material capital increases at a meaningful number of carriers, coordinated industry pushback is likely. The American Council of Life Insurers (ACLI) has historically engaged constructively with the GOES project, but a large negative capital impact at year-end 2027 would face significant resistance. Pushback could take the form of requesting transition relief, calling for formula factor recalibration alongside the generator change, or requesting an additional year of dual-run reporting before the AIRG retirement.

Operational Readiness at Carriers

Many life carriers run their VM-20 and VM-21 production processes on legacy actuarial software stacks that require vendor releases supporting GOES before companies can retire AIRG. The major actuarial software vendors have been working on GOES integration, but production-quality releases that support the full set of VM-20 and VM-21 calculations are not yet broadly available. If vendor releases slip, companies may not be operationally ready for a year-end 2027 transition.

If the timeline does slip, the practical consequence is a second year of dual-run calculations: companies maintaining both AIRG and GOES infrastructure simultaneously, doubling the operational lift on the appointed actuary team and the modeling staff. From running dual generators on production-scale infrastructure, the marginal cost of dual reporting is not trivial: scenario library generation, stochastic on stochastic projections for VM-21, and reconciliation of attribution differences all consume significant compute and analyst time.

What Appointed Actuaries Should Do Before the Field Test Closes

For companies that intend to participate in the Summer 2026 field test, the runway is short. Five steps deserve attention now.

Build the Dual-Run Infrastructure

If the company has not already built the capability to run AIRG and GOES scenarios in parallel through the existing VM-20 and VM-21 production pipelines, that work needs to start immediately. Vendor support for GOES is improving but not uniform, and companies relying on multiple vendors across their valuation systems will need to coordinate releases. Internal generators that produce GOES-equivalent paths can serve as a stopgap if vendor support is delayed.

Identify the Representative Block

The field test allows companies to submit a representative block rather than the full in-force. Selecting that block carefully matters: a block that captures the company's material C-3 exposure produces the most informative field test result, but a block that is too large creates operational drag that may compromise submission quality. The right answer depends on the company's modeling infrastructure and the urgency of the operational timeline.

Run Sensitivity Analyses Early

The field test submission will be more useful (both to the company and to the Working Group) if it includes sensitivity tests around key assumptions: lapse and behavior assumption recalibration to GOES scenarios, hedge program performance under GOES tail paths, and reinvestment strategy assumptions in extended high-rate environments. Running these sensitivities early surfaces issues that may require methodological adjustments before the September submission window closes.

Coordinate With Reinsurers

Companies with significant ceded VA or FIA business need to understand how their reinsurers are planning to incorporate the GOES transition. Reinsurance recapture triggers tied to capital ratios, treaty experience refunds, and retrocession arrangements all depend on the underlying capital framework. A GOES-driven capital change at the cedant can flow through to the reinsurer's net position, and treaty terms may need to be updated to reflect the new capital basis.

Engage With ACLI and Trade Associations

The field test will be followed by a comment period on aggregated results and any proposed factor recalibrations. Companies that engage early with ACLI working groups and other trade association forums can shape the industry response to the field test results. Companies that wait until the comment period to engage may find that the consensus industry position has already crystallized around recommendations that may not align with their own capital interests.

Why This Matters Beyond the Capital Number

The GOES transition is not just a recalibration exercise. It is the largest single change to the U.S. life insurance economic scenario generator since AIRG was first adopted, and the framework choices baked into GOES will shape capital, reserves, asset-liability management, and product design for at least a decade. Three implications extend beyond the immediate capital recalibration.

Asset-Liability Management Recalibration

Life insurer ALM frameworks rely on consistent economic scenario generators across reserves, capital, pricing, and risk management. A change to the generator at the regulatory level forces companies to reconsider whether their internal generators (used for pricing, hedging, and economic capital) should be aligned with GOES or maintained separately. Most companies will move toward alignment over time, but the transition creates short-term inconsistencies that complicate hedge program design and product pricing decisions.

Product Design Implications

The GOES tail behavior, particularly the fatter equity left tail and the sustained high-rate scenarios, will likely make certain product designs more capital-intensive than they appear under AIRG. Long-dated guaranteed living benefits with high benefit base growth rates may face higher capital under GOES than they do today, which could prompt design changes for new business. Conversely, FIA designs that insulate against extended low-rate environments may see slightly lower capital under GOES because the AIRG bias toward low-rate paths is reduced.

Coordination With International Frameworks

The U.S. life RBC framework operates within a global regulatory environment that increasingly relies on integrated economic scenario generators. The IAIS Insurance Capital Standard (ICS), Solvency II internal models, and the Bermuda Economic Balance Sheet all use stochastic scenario frameworks for various calculations. Convergence between the U.S. GOES and international approaches is not a stated goal, but the technical similarity of the GOES design to other modern economic scenario generators makes future cross-jurisdictional coordination more feasible.

The Bottom Line

The Summer 2026 C-3 field test is the most consequential life RBC exercise in at least a decade. The GOES transition fundamentally changes the stochastic foundation for variable annuity and fixed indexed annuity capital, and the parallel CLO RBC factor track simultaneously affects the C-1 component for the same companies. The year-end 2027 adoption target is achievable but not assured: vendor readiness, calibration concerns surfaced by the field test, and coordinated industry pushback could each push the timeline to year-end 2028.

For appointed actuaries and chief actuaries at U.S. life insurers writing material C-3 business, the field test is the last meaningful opportunity to influence the GOES design and the C-3 formula factors before adoption. Participation in the field test is not just a regulatory exercise but a strategic opportunity to ensure that the final framework reflects the operational realities of running large blocks of long-dated business under stochastic capital regimes. Companies that engage constructively, submit complete and well-documented field test results, and participate actively in the post-submission comment process will help shape the framework that they and their competitors will operate under for the next decade.

This continues a pattern visible across the NAIC's recent capital and valuation modernization work. Major changes to the life regulatory framework are increasingly developed through field tests and parallel-run exercises rather than through pure top-down rulemaking, which gives industry both more influence and more responsibility in shaping the final outcome. The actuaries who treat the field test as a serious technical and strategic exercise, rather than a compliance task, will be best positioned for the year-end 2027 transition.